A Universal Mobile Telecommunication System (UMTS) is a 3G wireless communication network standard defined by the Third Generation Partnership Project (3GPP). A UMTS network consists of a Core Network (CN) and an Access Network (AN). The CN consists of a Circuit Switched (CS) domain and a Packet Switched (PS) domain. The CS domain provides CS-based services such as voice services. The PS domain provides PS-based services such as Internet access. A terminal used by a mobile subscriber is a User Equipment (UE).
In a prior art UMTS network, the CN generally includes a PS domain and a CS domain. The PS domain in the CN generally includes a Serving GPRS Support Node (SGSN), a Gateway GPRS Support Node (GGSN), and a Home Location Register (HLR). The CS domain on the CN generally includes a Mobile Switching Center (MSC), a Visitor Location Register (VLR), and a Gateway Mobile Switching Center (GMSC). The AN generally includes a Radio Network Controller (RNC) and a NodeB. Each RNC is connected to several NodeBs. Each SGSN is connected to several RNCs. An Iu interface is a key interface between the AN and the CN. Management and control of radio resources are isolated on the AN by the Iu interface, and thus the CN focuses solely on the service provision.
On a traditional network such as an R99 system, one RNC is connected to only one CN node. For example, one RNC is connected to only one SGSN. Thus, problems such as single point failure exist. If an SGSN is down, a UE in the service area of the SGSN cannot access the network, and thus cannot perform communications. To rectify the foregoing defect, a concept of “flex” is introduced. When a many-to-many relationship between AN devices and CN devices exists on an Iu interface, the Iu interface is called Iu-flex. In FIG. 1, there are Areas 1 to 8 governed respectively by 8 RAN nodes. A configuration of a “pool area” in the prior art is shown herein. In FIG. 1, a Radio Access Network (RAN) node (for example: an RAN node may be an RNC in legacy wireless network such as an R99 system) is connected to multiple SGSNs or multiple MSCs, and an SGSN or an MSC is connected to multiple RAN nodes. These SGSNs form a pool, and the areas governed by the RAN nodes which the SGSNs connect are called pool area (for example: CS pool—area 1 including Areas 1, 2, 5 and 6, CS pool—area 2 including Areas 2, 3, 6 and 7, PS pool—area 1 including Areas 1 and 5, and PS pool—area 2 including Areas 2, 3, 6 and 7,). In a pool, multiple CN nodes such as SGSNs are connected to all Radio Access Network (RAN) nodes (such as RNCs) in the pool (for example: SGSNs 3 to 5 are connected to each RAN node in PS pool—area 2 and SGSNs 1 to 2 are connected to each RAN node in PS pool—area 1), which is different from a traditional mode where one AN node is connected to only one CN node. When a UE enters a pool area initially, an RAN node can select one CN node according to load sharing principles. Thus, when the UE moves or accesses the network in the pool, the UE is always anchored at the selected CN node. Therefore, single point failure and frequent relocation of CN nodes can be prevented because the UE does not need to change the CN node in the pool.
In the prior art, a network allocates a Temporary Mobile Station Identity (TMSI, which is allocated by an MSC in a CS domain) or a P-TMSI (which is allocated by an SGSN in a PS domain) to a UE after the UE is attached to the network.
FIG. 2 shows a network with TMSI/P-TMSI design in a prior art. The network includes four pools (Pools 1 to 4) whose areas have overlapped parts. Each pool includes five CN devices, which are differentiated with different Network Resource Identifiers (NRIs) (Pool 1 with NRIs 1 to 5, Pool 2 with NRIs 16 to 20, Pool 3 with NRIs 11 to 15, Pool 4 with NRIs 6 to 10). A Non Access Stratum (NAS) Node Selection Function (NNSF) and the TMSI uniqueness of a UE in a paging area are not affected, and therefore, duplicate NRIs can be used in non-adjacent pool areas (for example, NRI 11 can be reused in a pool area not adjacent to area related to pool 3, NRI 1 can be reused in a pool area not adjacent to area related to pool 1). Assume that each CN device can attach a maximum of 1,000,000 subscribers, while the overlapped pool areas have 12,000,000 subscribers, and other areas have fewer subscribers.
On the preceding network, 20 CN devices are sufficient to attach 12,000,000 subscribers. An NRI may be set to 5 bits (25=32, which can be used to identify 20 CN devices). The independently allocated ID of each device is 21 bits (1000000=220, which can be used to identify 2,000,000 subscribers), two bits are used to differentiate a PS domain from a CS domain, and the remaining four (32−5−21−2=4) bits are used for restart.
FIG. 3 shows a structure of the flex on an SAE network in the prior art. In a pool (such as Mobility Management Entity (MME) pool 1 or MME pool 2), multiple CN nodes such as MMEs (not shown in the figure) are connected to all RAN nodes such as eNodeBs (ENBs) 1 to 4, which is similar to the method of the prior art. When a UE initially enters the area related to the pool, an RAN node can select one CN node according to load sharing principles. Thus, when the UE moves or accesses the network in this pool, the UE is always anchored at the selected CN node. On the SAE network, pool areas may also be overlapped. In addition, the SAE network specifies that an MME pool area or an S-GW pool area includes a complete Tracking Area (TA, which is similar to a Location Area (LA) or a Routing Area (RA) on a UMTS network).
Assume that a UE is allocated one TA once. When the UE enters MME pool 1 (hereinafter referred to as MP1) for the first time, for example, when the UE enters ENB1, the UE selects one MME from MP1. When the UE moves from ENB1 to ENB2 and then to ENB3, the UE does not need to change the MME. When the UE moves to ENB4 that is not connected to the source MME and belongs only to MP2, the UE may reselect an MME in MP2. In FIG. 3, ENB2 and ENB3 belong to two MME pools; that is, ENB2 and ENB3 are connected to all MMEs in the two MME pools. Therefore, ENB2 and ENB3 are overlapped parts of MP1 and MP2. The advantage of overlapping is as follows: because ENB3 is connected to MP2, when the UE returns from ENB4 to ENB3, the UE does not need to reselect an MME until the UE returns to ENB1, thus preventing ping-pong MME relocation. If ENB3 is not connected to MP2, ping-pong MME relocation occurs when the UE moves between ENB3 and ENB4.
With respect to the TA concept, it should be noted that on an SAE network, multiple TAs can be allocated to a UE, which is different from the practice in a UMTS network where only one LA or RA can be allocated to a UE. In FIG. 3, if a UE is registered in a pool, and if the TA list includes TA1 and TA2, the UE does not need to initiate an update when moving between ENB1 and ENB2. That is, the UE does not need to initiate an update when moving in the allocated TA list.
Currently, a TMSI problem about the SAE is as follows: the SAE needs to support multiple Radio Access Technologies (RATs), and therefore, several types of terminals access the network; to keep the system capacity, an SAE-TMSI (namely, SAE TMSI) may need to be extended to support more subscribers, to expand the capacity, and to simplify the network.
When assessing the operation of prior art networks the inventor identified at least the following problem: when a UE moves from an SAE network to a legacy (such as 2G/3G) network, the legacy network cannot identify the SAE-TMSI, and therefore, the newly selected SGSN cannot find the originally-assigned MME on the SAE network. As a result, the new SGSN cannot obtain the context of the UE. And, ongoing services for the UE maybe interrupted.